The newest generation of space explorers has some unusual characteristics. In addition to four-wheel mobility and the ability to withstand extremes of temperature, the most desirable quality for a new rover is autonomous operation versus the remote control that is currently used with rovers. For example, NASA’s Mars rovers Spirit and Opportunity are tele-operated by humans on Earth, and with signals needing to travel 140 million miles between Earth and Mars, even simple actions are difficult to coordinate.
For a rover to be autonomous on Mars, however, the issue of navigation gets tricky: Earth-referenced sensors – GPS and other magnetic heading sensors – are useless for navigating on Mars, which does not have the satellite infrastructure for its own GPS system (although Mars GPS has been discussed). For autonomous space rovers, Earth-based devices such as GPS and compasses are being replaced with inertial sensors that are truly out of this world.
Teams of engineers are currently working on the parts and programming needed to create autonomous space rovers, whose tasks would include seeking out samples and returning to a designated point. Spurring the research along are competitions such as the Sample Return Robot Challenge, jointly sponsored by Worcester Polytechnic Institute (WPI) and NASA.
The most recent challenge was held in June on the campus of WPI, and more than 10 teams participated by putting their robotic rovers to the test, including the Robotics and Intelligent Vehicles Research Laboratory (RIVeR) at WPI, which entered AERO (Autonomous Exploration Rover), in the contest.
The student-designed AERO includes such advanced elements as a differential-drive four-wheeled mobility platform and a six-degrees-of-freedom manipulator, a stereo vision system, LIDAR, and an inertial measurement unit (IMU). To perform the sample retrieval, AERO needs to process data that enables it to not only avoid obstacles and navigate to its goal, but also locate and identify the samples it should collect. For the sample identification and retrieval tasks, AERO uses stereo vision object recognition, localization, and grasping algorithms to control the manipulator and retrieve the samples.
The primary navigation sensor on AERO is a KVH 1750 IMU, a fiber optic gyro (FOG)-based inertial measurement unit integrated with AERO’s odometer to provide dead-reckoning navigation. AERO, with KVH’s 1750 IMU, maps its environment, locates itself on that map, and then searches for samples to retrieve.
“We’re excited about our involvement,” says Jay Napoli, KVH Vice-President, FOG & OEM Sales, who says that the knowledge gained from the Sample Return Challenge could contribute to other cutting-edge technology, including driverless cars.
Autonomous rovers could also prove useful in dangerous situations close to home – for example, monitoring earthquakes and volcanoes, or handling bio-hazardous material.
For now, AERO is a very busy rover here on Earth: The robot was a crowd pleaser at the recent show of the Association for Unmanned Vehicle Systems International (AUVSI) in Washington, DC. A technical paper about AERO, written by the WPI team, has been accepted for the upcoming 2013 IEEE Systems, Man, and Cybernetics (SMC) Conference in October. In addition, the 2014 Sample Robot Challenge was recently announced; it will be held June 9-14, 2014, at WPI.